Contrast Calculator

Written in Java, the partially automated contrast calculator aims to simplify the process of finding the high and low values needed to calculate contrast. It runs faster than the IMAQ Vision Builder program because it is dedicated to one purpose and therefore does not have all of the overhead like the other program.

It places the histogram, data output, and image selection all on one screen so multiple readings can be taken more rapidly. The histogram has also been improved by making it larger and easier to read with a zoom function which expands the x-axis. It also includes a feature which allows the user to enter high and low values into two different forms to computer Michelson contrast for both monochrome and color screens. Currently the program only supports JPEG image files, but a less stable, experimental version also supports TIFFs. There is almost no variation between the outputs using the two different image files. The problem with loading TIFFs is the immense amount of memory required to load the image. JPEGs require approximately 56 MB of memory while TIFFs require anywhere from 300-700 MB of memory. To circumvent some of the overhead, the program currently loads a JPEG into the display and allows the user to select the area of the image to be measured, and once the selection has been made, it takes the measurements from a TIFF version of the image. This method runs faster, but not as fast as using just the JPEG. Another problem with this is the file must be available in both TIFF and JPEG for the program to read, so the .CR2 files must be converted into both and held within the same folder.

Finding Light Values in Color Screens

For monochrome screens, the light and dark peaks are both distinct single peaks making it easy to choose a point as the light and dark values for Michelson contrast. The dark peak in colored screens is usually easy to find, but the light peak is at best a triplet of peaks, and at worst an amorphous blob that is a combination of all three peaks on top of one another such as in the following image:

To solve this problem, we added a filter to single out the individual red, green, and blue peaks. For example, to find the red peaks in the image, the program simply eliminates all RGB values in which red is not the largest value.

This results in a much cleaner looking graph which contains more useful information than the combined histogram. From this we extract the high peaks of the red, green, and blue values and enter them into the formula (0.2126*red + 0.7152*green + 0.0722*blue) to find the luminance of the light peak for the screen. Now that both the light and dark peaks of the color screen have been calculated, we calculate Michelson contrast using the same method as with monochrome screens. We would like to find a way to implement an algorithm that finds the average of each peak instead of just finding the absolute maximum value.

American Association of Physicists in Medicine (AAPM) Task Group 18

The American Association of Physicists in Medicine (AAPM) Task Group 18 is a national task force consisting of medical imaging experts and organizational affiliates dealing with performance evaluation of electronic display devices. The purpose of the Task Group is to generate a document that provides guidelines to practicing medical physicists and engineers for in-field performance evaluation of electronic display devices intended for medical use. Some of the goals of the Task Group include:

1. Standardization: providing standard testing methods for evaluating the performance of electronic display devices utilized in medicine, allowing inter- and intra-institutional comparisons
   
2. Establishing performance criteria: recommending minimum performance requirements for safe utilization of electronic displays in medical and radiologic applications
   
3. Education: educating medical physicists and other health-care professionals about display technology and display quality issues relevant to medical imaging
   
4. Communication: facilitating communication between industry, medical physicists, researchers, and other individuals/organizations dealing with display quality characterization
   
5. Professional: further expanding the role of medical physics in the growing area of PACS
   

The AAPM TG18's excellent report and valuable guide on the "Assessment of Display Performance for Medical Imaging Systems" can be found here (PDF, 12MB).

Further information about the AAPM TG18, links to downloadable test patterns and other valuable links can be found here.

Color Usage Research Lab @ NASA Ames

Here's a very useful site at NASA Ames Research Center on "Using Color in Information Display Graphics"